UPANets: Learning from the Universal Pixel Attention Neworks
With the successful development in computer vision, building a deep convolutional neural network (CNNs) has been mainstream, considering the character of shared parameters in a convolutional layer. Stacking convolutional layers into a deep structure improves performance, but over-stacking also ramps...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-09-01
|
| Series: | Entropy |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1099-4300/24/9/1243 |